SEAS 2026: 1st Frontier Workshop on Self-Evolving AI Safety
SEAS 2026 will be held in the Cross-AI Symposium 2026.
Introduction
Artificial Intelligence (AI) is rapidly advancing toward increasingly autonomous agents capable of long-horizon reasoning, tool use, adaptation, self-improvement, and coordination with other agents. These capabilities create transformative opportunities while raising a fundamental safety challenge: our ability to understand, evaluate, and control AI systems must keep pace with their increasing autonomy and capability.
This challenge is especially acute for self-evolving, self-recursive, and collective AI agents. Agents that can modify their operation, acquire new capabilities, or coordinate with other agents may exhibit emergent behaviors, unexpected interactions, and capability growth that are difficult to anticipate or control. Recent incidents involving frontier agents crossing intended security or authorization boundaries further demonstrate the need for effective mechanisms to prevent unauthorized actions, enforce safety constraints, detect failures, and enable reliable intervention.
The central premise of this workshop is that AI safety must advance alongside AI capability. Increasingly powerful agents must remain controllable, secure, verifiable, and accountable to humans, including as they adapt, interact, and evolve. Achieving this requires connecting fundamental scientific advances with practical engineering, deployment experience, and effective governance.
The workshop brings together academia, industry, and government to address frontier AI safety as a shared scientific, technical, operational, and societal challenge. We welcome contributions spanning both fundamental research and practical solutions, with emphasis on approaches that translate research into measurable, deployable, and enforceable safety mechanisms for increasingly autonomous agents.
Topics
This open call for papers invites original contributions on the theory, methods, architectures, evaluation, and practical deployment of safe agentic systems. Submissions may address technical, conceptual, empirical, or applied research, including, but not limited to, the following areas:
- Self-Evolving AI: Self-modification, recursive improvement, capability growth, and safety under evolution.
- Collective Intelligence: Multi-agent coordination, emergent capabilities, communication, and collective behavior.
- Control & Alignment: Human control, goal robustness, corrigibility, interruptibility, and alignment under capability change.
- Agent Security & Containment: Unauthorized actions, privilege escalation, sandboxing, isolation, and secure shutdown.
- Safety Evaluation: Benchmarks, red teaming, adversarial testing, continuous evaluation, and emergent-capability detection.
- Oversight & Verification: Scalable oversight, runtime monitoring, formal verification, safety assurance, and intervention.
- Safe Deployment & Governance: Deployment controls, incident response, standards, accountability, and public-sector applications.
- Research-to-Practice: Translating scientific advances into practical safety engineering, operational safeguards, and deployable solutions across academia, industry, and government.
Important Dates (AoE)
| Milestone | Cross-AI Archival Paper Deadline | Non-Archival Paper Deadline |
|---|---|---|
| Full/Short Paper Submission | October 26, 2026 | N/A |
| Poster/Demo Paper Submission | November 2, 2026 | November 16, 2026 |
| Abstract Submission | November 2, 2026 | November 16, 2026 |
| Acceptance Notification | November 16, 2026 | November 30, 2026 |
| Camera-Ready Submission | November 30, 2026 | N/A |
| Pre-Recorded Video Submission | November 30, 2026 | November 30, 2026 |
| Author Registration Deadline | December 2, 2026 | December 2, 2026 |
| Cross-AI Symposium 2026 | December 14-16, 2026 |
Archival work submissions for the Cross-AI Symposium must be submitted by the first deadline. Non-archival submissions may be submitted by either deadline.
Submission
All submissions are made through the AirBalloon conference management system and evaluated through a unified process.
1. Unpublished and Unreviewed Papers
Unpublished and unreviewed papers should follow the IEEE conference manuscript template (Overleaf or US Letter), be written in English, submitted as PDF files, and undergo double-blind review.
| Category | Length (including references) |
|---|---|
| Full paper | 8–10 pages |
| Short paper | 5–7 pages |
| Poster paper | 3–4 pages |
| Demo paper | 2–3 pages with a link to a 2-minute demo video |
| Extended Abstract | 1 page |
2. Reviewed but Not Yet Published Papers
For papers previously submitted to major AI/ML/NLP/Computer Vision/Robotics/Signal Processing/Graphics conferences (e.g., NeurIPS, ICML, ICLR, ACL, EMNLP, CVPR, ECCV, AAAI, ICRA, IROS, CoRL, SIGGRAPH, ACM MM, ICASSP, HCI, KDD), please include the original reviews as supplementary material, clearly indicating the conference, along with a separate document summarizing the revisions addressing the reviewers’ comments. This may help expedite the review process.
3. Non-Archival Submissions
Non-archival submissions may include published journal or conference papers, as well as works in progress seeking feedback. All are welcome for presentation at Cross-AI events. For published papers, authors only need to submit an abstract and the original DOI link.
4. E-Posters and Demo Videos
Electronic posters and/or demo videos are optional for all submission types (Full, Short, Poster, Demo, and Non-Archival).
| Category | Description |
|---|---|
| E-Poster | 1-page electronic poster (PDF) |
| Demo Video | up to 2-minute demo video |
5. Pre-Recorded Presentation Videos
Accepted submissions provide a pre-recorded video.
| Category | Presentation | Q&A |
|---|---|---|
| Full paper | 15 minutes | 3 minutes |
| Short paper | 10 minutes | 2 minutes |
| Poster | 6 minutes | 2 minutes |
| Demo | 5 minutes (system description + showcase) | 2 minutes |
| Abstract | 5 minutes | 2 minutes |
Publishing & Registration Requirements
All accepted archival papers will be published in the indexed Cross-AI conference proceedings. At least one author must register for the workshop. Authors may optionally have their papers considered for further publication in an IEEE-published Cross-AI Special Edition in 2027, subject to Cross-AI approval and an additional publication fee.
Media Hosting: Submitted presentation and demo videos, e-posters, audio, and other approved media will be hosted on the conference platform and made openly accessible for three years. Media hosting and open access are included in the Author registration fee.
Cross-AI Group
This group serves as a forum for notices and announcements of interest to the Cross-AI community. This includes news, events, calls for papers, calls for collaborations between academia and industry, dataset releases, employment-related announcements, etc.
Welcome to subscribe to the Cross-AI group.
If you are interested in serving on the workshop program committee or paper reviewing, please apply online.